Nested For Loop; Combine Two for loops

Hi, time is an important factor so I appreciate any help soon. Thank you!

I am writing code to identify two populations of cells with varying sigma, mu, and quanitities. So far, I am varying only the sigma_sub of the sub (smaller) population, while keeping the other variables constant. The way the code works is there is a for loop that iterates through a set of sigma_sub pre defined values, picks one postion of the iteration and sets sigma_sub to that value. Then, stores this value in an array through the length of the predefined values.

%% for loop for sigma sub values
% iterates through predefined values, picks position, assigns sigma value
sigmasub_val = 0.1:0.1:3;
a_sigmasub=[];
for i = 1:length(sigmasub_val)
    sigmasub_pos = randi(length(sigmasub_val));
    sigma_sub = sigmasub_val(sigmasub_pos);
    a_sigmasub =[a_sigmasub;sigma_sub];
end

Next, this chaging value and the constant variables are used to find a model that best represents the data. The other for loop runs 4 tmes through different models to find the best one, and outputs the value of the numComponents of the best model for the given sigma_sub value and constants.

y = sigma_main.*randn(n_main,1) + mu_main;    %10^6 SKBr3 cells
y2 = sigma_sub.*randn(n_sub,1) + mu_sub; %100,000 MDA MB 231
C = cat(1, y, y2);
AIC = zeros(1,4);
GMModels = cell(1,4);
options = statset('MaxIter',00);
for k = 1:4
    GMModels{k} = fitgmdist(C,k);
    AIC(k)= GMModels{k}.AIC;
end

[minAIC,numComponents] = min(AIC);
numComponents;

I need to find a way to combine this. So for every value of sigma_sub, have 4 models be tested on each value, and output the best model.

Does this make sense? Please help!

Answer 1

[Geoff Hayes]

Elizabeth - perhaps you can combine the two as follows

sigmasub_val = 0.1:0.1:3;
outputData = zeros(length(sigmasub_val), 2);    % <--- create an output array for sigmasub,numComponents
for i = 1:length(sigmasub_val)
    sigmasub_pos = randi(length(sigmasub_val));
    sigma_sub = sigmasub_val(sigmasub_pos);
    
    y = sigma_main.*randn(n_main,1) + mu_main;    %10^6 SKBr3 cells
    y2 = sigma_sub.*randn(n_sub,1) + mu_sub; %100,000 MDA MB 231
    C = cat(1, y, y2);
    AIC = zeros(1,4);
    GMModels = cell(1,4);
    options = statset('MaxIter',00);
    for k = 1:4
        GMModels{k} = fitgmdist(C,k);
        AIC(k)= GMModels{k}.AIC;
    end
    [minAIC,numComponents] = min(AIC);

    outputData(i,1) = sigma_sub;
    outputData(i,2) = numComponents;
    
end

Is that something close to what you are looking for?

[Elizabeth Cardona]

This is it, exactly! Thank you. It just takes a very very long time to run.